Maybe you need to put an A/D convertor in the tonearm and pass the signal digitally to the next stage, that should eliminate the hum problem.
JR
JR
Phono Preamp Grounding Behaves like an Antenna...
- Thread starter thermionic
- Start date
Help Support GroupDIY Audio Forum:
bcarso
Well-known member
Optical, clearly; power the preamp and A/D with a laser and photovoltaic stack via low-loss fiber, and send the bitstream back over another fiber.
Now if that torques the arm too much, then work out a direct optical system. That would have the added advantage of adjustment of antiskating with radiation pressure, and would provide a built-in cartridge protection system for audiophiles with kids.
"Jimmy, stop whining. Daddy TOLD you not to touch it, and now you'll have a scar there for the rest of your life. Maybe next time you'll LISTEN!!"
:razz:
Now if that torques the arm too much, then work out a direct optical system. That would have the added advantage of adjustment of antiskating with radiation pressure, and would provide a built-in cartridge protection system for audiophiles with kids.
"Jimmy, stop whining. Daddy TOLD you not to touch it, and now you'll have a scar there for the rest of your life. Maybe next time you'll LISTEN!!"
:razz:
thermionic
Well-known member
- Joined
- Jun 3, 2004
- Messages
- 1,671
thermionic
Well-known member
- Joined
- Jun 3, 2004
- Messages
- 1,671
Hi,
Before I go the “virtual ground” route, I’d like to clarify the exact issue and see if I’m missing something.
This is basically what I’m using: http://www.dself.dsl.pipex.com/ampins/discrete/dsctamp2.gif
My problem is finding a place to ground the resistor and cap that come off the back of Q3’s base to ground (R6 / 47uF – (1,000uF in my case)).
I have established:
If I ground the cap to the op-amp’s ground leg (my local “star”), all hell breaks loose.
If I ground the cap before all the other grounds, right at the power socket or battery virtual ground, leaving the rest of the grounds in a star config at op-amp GND pin, I get very good performance.
The cap likes the first ground point all to itself; if I ground the caps from other op-amps, i.e. I try and power more than one op-amp off the same ground rail, things get noisy, like an antenna. The more op-amps return to this point, the more noise.
If I introduce a microscopic amount of resistance between power ground and the cap GND, it gets noisy: I have a thick copper bar from DC power socket for distribution. The further cap is grounded along bar, the more noise I get.
Grounding the cap at a point that has zero resistance to signal input GND seems to cause a loop. The cap wants minimal impedance / resistance to power ground, but before all the other grounds get their tap.
I have tried situating the regulated supply right next to preamp and jumping a 30A cable from CT / filter-cap centre-point to reduce impedance to ground. This makes no difference at all. The length of ground lead from PSU to preamp makes no difference… The gauge of GND cable in preamp makes no difference.
I can get one channel to perform beautifully; the noise adds up incrementally the more preamp channels I power up.
Please note that the problem is the same whether I use shorting plugs or a cartridge for the input. If anything, it's slightly quieter with a cart as it puts more resistance between input and GND.
Thanks again :guinness:
Olaf,
If your problem is anywhere near as difficult to solve as mine, I would say that you need to make a duplicate power-supply for your headphone amp and go full dual-mono. Don’t forget to break out individual ground leads for each headphone speaker!
Justin
Before I go the “virtual ground” route, I’d like to clarify the exact issue and see if I’m missing something.
This is basically what I’m using: http://www.dself.dsl.pipex.com/ampins/discrete/dsctamp2.gif
My problem is finding a place to ground the resistor and cap that come off the back of Q3’s base to ground (R6 / 47uF – (1,000uF in my case)).
I have established:
If I ground the cap to the op-amp’s ground leg (my local “star”), all hell breaks loose.
If I ground the cap before all the other grounds, right at the power socket or battery virtual ground, leaving the rest of the grounds in a star config at op-amp GND pin, I get very good performance.
The cap likes the first ground point all to itself; if I ground the caps from other op-amps, i.e. I try and power more than one op-amp off the same ground rail, things get noisy, like an antenna. The more op-amps return to this point, the more noise.
If I introduce a microscopic amount of resistance between power ground and the cap GND, it gets noisy: I have a thick copper bar from DC power socket for distribution. The further cap is grounded along bar, the more noise I get.
Grounding the cap at a point that has zero resistance to signal input GND seems to cause a loop. The cap wants minimal impedance / resistance to power ground, but before all the other grounds get their tap.
I have tried situating the regulated supply right next to preamp and jumping a 30A cable from CT / filter-cap centre-point to reduce impedance to ground. This makes no difference at all. The length of ground lead from PSU to preamp makes no difference… The gauge of GND cable in preamp makes no difference.
I can get one channel to perform beautifully; the noise adds up incrementally the more preamp channels I power up.
Please note that the problem is the same whether I use shorting plugs or a cartridge for the input. If anything, it's slightly quieter with a cart as it puts more resistance between input and GND.
Thanks again :guinness:
Olaf,
If your problem is anywhere near as difficult to solve as mine, I would say that you need to make a duplicate power-supply for your headphone amp and go full dual-mono. Don’t forget to break out individual ground leads for each headphone speaker!
Justin
OK, let me try saying this another way. Yes, indeed if you just run one ground wire between three different turntables and preamps, all connected together, the voltage differences in that single conductor can cause significant problems.
Attempting to find the "magic" quiet point to connect sundry circuits is one part wishful thinking and a rejection of basic engineering principles. Your difficulty trying to do this is not surprising. It is very likely impossible to do, and quiet today, may be noisy tomorrow if some external factor changes.
There is a method to manage different ground potentials, namely use of differential circuitry. A signal is treated as (signal +)-(signal -). This maintains signal integrity. Note ground can not be hard bonded to signal- at both ends of a span.
Warning- I'm about to repeat myself again, but I'll change it a little to keep it interesting for myself.
Instead of thinking of one ground, lets give "them" cute names; RIAA gnd 1,2,3, chassis/ps gnd, turntable gnd 1,2,3. And non grounds, Cart L, R, +, -, 1,2,3.
The bottom of your 1000uf gain cap in preamp 1 connects hard to RIAA gnd 1, 2 to 2, 3 to 3. The phono cart, L & R - also connect to these RIAA gnds respectively (but not each other.
We will now have clean uncorrupted RIAA equalized signals at the output of these 3 different RIAA stages, but this clean signal is relative to or reference to these 3 RIAA gnds. The task now is how to reference these 3 grounds to a common PS gnd, chassis ground, and to following equipment. It is possible that you might be able to hard bond these 3 RIAA gnds directly to the PS gnd, but I would be nervous about potential loops so would connect through a small compliance (say 10 ohms and .1uF).
Now the final remaining issue is getting the clean signal out of your box. If the following equipment is balanced or true differential input (+,-,and gnd) you could simply connect output + to each RIAA output, output - to each RIAA gnd, and output gnd to PS/chassis ground. For stability and balance I would use 50-100 ohms build out resistors in series with each + and - output.
If you need to output a 2 circuit unbalanced signal referenced to a common chassis ground you will need 3 more differential amplifiers (1 each opamp w/4 resistors) to reference between the RIAA gnds and the common output ground (or perhaps the following product's unbalanced input -/gnd).
Final housekeeping is that turntables should be hard grounded to chassis ground and hopefully turntable gnds are isolated from tone arm wiring.
I hope this finally makes sense. If you think it's difficult making three phonos play nice together, try putting 40 mic preamps in a single chassis..
JR
Note: in a single preamp it is probably possible to bond the RIAA gnd to the chassis gnd at the unbalanced output and get away with it. The combination of 3 different preamps confounds this simpler grounding scheme.
Attempting to find the "magic" quiet point to connect sundry circuits is one part wishful thinking and a rejection of basic engineering principles. Your difficulty trying to do this is not surprising. It is very likely impossible to do, and quiet today, may be noisy tomorrow if some external factor changes.
There is a method to manage different ground potentials, namely use of differential circuitry. A signal is treated as (signal +)-(signal -). This maintains signal integrity. Note ground can not be hard bonded to signal- at both ends of a span.
Warning- I'm about to repeat myself again, but I'll change it a little to keep it interesting for myself.
Instead of thinking of one ground, lets give "them" cute names; RIAA gnd 1,2,3, chassis/ps gnd, turntable gnd 1,2,3. And non grounds, Cart L, R, +, -, 1,2,3.
The bottom of your 1000uf gain cap in preamp 1 connects hard to RIAA gnd 1, 2 to 2, 3 to 3. The phono cart, L & R - also connect to these RIAA gnds respectively (but not each other.
We will now have clean uncorrupted RIAA equalized signals at the output of these 3 different RIAA stages, but this clean signal is relative to or reference to these 3 RIAA gnds. The task now is how to reference these 3 grounds to a common PS gnd, chassis ground, and to following equipment. It is possible that you might be able to hard bond these 3 RIAA gnds directly to the PS gnd, but I would be nervous about potential loops so would connect through a small compliance (say 10 ohms and .1uF).
Now the final remaining issue is getting the clean signal out of your box. If the following equipment is balanced or true differential input (+,-,and gnd) you could simply connect output + to each RIAA output, output - to each RIAA gnd, and output gnd to PS/chassis ground. For stability and balance I would use 50-100 ohms build out resistors in series with each + and - output.
If you need to output a 2 circuit unbalanced signal referenced to a common chassis ground you will need 3 more differential amplifiers (1 each opamp w/4 resistors) to reference between the RIAA gnds and the common output ground (or perhaps the following product's unbalanced input -/gnd).
Final housekeeping is that turntables should be hard grounded to chassis ground and hopefully turntable gnds are isolated from tone arm wiring.
I hope this finally makes sense. If you think it's difficult making three phonos play nice together, try putting 40 mic preamps in a single chassis..
JR
Note: in a single preamp it is probably possible to bond the RIAA gnd to the chassis gnd at the unbalanced output and get away with it. The combination of 3 different preamps confounds this simpler grounding scheme.
funkydiplomat
Well-known member
yes. maybe check/add decoupling at DC power inlet.
thermionic
Well-known member
- Joined
- Jun 3, 2004
- Messages
- 1,671
Although I was intending to thank John for taking the time to write some great posts, part of me was hoping this thread would drift off the bottom of the page...
It transpires that the issue was due to intermittent oscillation, not grounding practice per-se.
I can rule out ground issues as I managed to trigger the oscillation on the bench, running via batteries, into output transformers. The preamp was fine, with just a touch of mains-freq-related ripple (all phono preamps have a touch of it – less than 3mV).
As soon as I connected something to the secondary of the transformer, I noticed something whiz past on the ‘scope that I hadn’t seen before… When I clicked the ‘scope frequency control all the way to the right (it’s a 20Mhz ‘scope – my 50Mhz needs repairing), I could see a vicious oscillation…
The funny thing is that the oscillation doesn’t start until you load the output… I was hearing the RFI and LF hum that accompanies oscillation, but was unaware of the – AS PER F*CKING USUAL – “elephant” in the room with me…
I now have 2 channels running perfectly on the bench with no problems.
My gut instinct tells me that the oscillation is to do with power-supply output impedance, which would explain why it gets ****** when you connect extra preamps. I can’t see anything on the power rails at 1mV p / div with 6 preamps in place, so I don’t think it’s due to ripple under load.
Can anyone give a reason why PSU output impedance would make a design prone to oscillation? It’s a single-ended op-amp btw…
Today, I will be trying a lot of additional PSU decoupling (as you suggest, Diplomat – thanks!) per preamp. I hope that 1,000u (63v) per rail, per op-amp will do the trick.
It was good-as-gold last night for 2 channels… Can I extend this to extra channels???
The guy who gave me the design is seriously pissed off with me btw.
Thanks again,
Justin
It transpires that the issue was due to intermittent oscillation, not grounding practice per-se.
I can rule out ground issues as I managed to trigger the oscillation on the bench, running via batteries, into output transformers. The preamp was fine, with just a touch of mains-freq-related ripple (all phono preamps have a touch of it – less than 3mV).
As soon as I connected something to the secondary of the transformer, I noticed something whiz past on the ‘scope that I hadn’t seen before… When I clicked the ‘scope frequency control all the way to the right (it’s a 20Mhz ‘scope – my 50Mhz needs repairing), I could see a vicious oscillation…
The funny thing is that the oscillation doesn’t start until you load the output… I was hearing the RFI and LF hum that accompanies oscillation, but was unaware of the – AS PER F*CKING USUAL – “elephant” in the room with me…
I now have 2 channels running perfectly on the bench with no problems.
My gut instinct tells me that the oscillation is to do with power-supply output impedance, which would explain why it gets ****** when you connect extra preamps. I can’t see anything on the power rails at 1mV p / div with 6 preamps in place, so I don’t think it’s due to ripple under load.
Can anyone give a reason why PSU output impedance would make a design prone to oscillation? It’s a single-ended op-amp btw…
Today, I will be trying a lot of additional PSU decoupling (as you suggest, Diplomat – thanks!) per preamp. I hope that 1,000u (63v) per rail, per op-amp will do the trick.
It was good-as-gold last night for 2 channels… Can I extend this to extra channels???
The guy who gave me the design is seriously pissed off with me btw.
Thanks again,
Justin
bcarso
Well-known member
[quote author="thermionic"]
The guy who gave me the design is seriously pissed off with me btw.
[/quote]
He'll get over it.
The guy who gave me the design is seriously pissed off with me btw.
[/quote]
He'll get over it.
Buried in my advice was to use build out impedances in series with output.
"For stability and balance I would use 50-100 ohms build out resistors in series with each + and - output."
How much impedance are you using in series with your outputs?
I have encountered PS oscillations when there was a significant compliance between the power supply ground and the load ground for that PS, but if this is only at VHF you should be able to manage it with appropriate build outs, and perhaps a few disc caps.
JR
"For stability and balance I would use 50-100 ohms build out resistors in series with each + and - output."
How much impedance are you using in series with your outputs?
I have encountered PS oscillations when there was a significant compliance between the power supply ground and the load ground for that PS, but if this is only at VHF you should be able to manage it with appropriate build outs, and perhaps a few disc caps.
JR
